Run threat assessments through multiple AI models and log where they disagree....
http://www.video-bookmark.com/user/lucyreid88
Run threat assessments through multiple AI models and log where they disagree. Write down every assumption, name the person who makes the final call, and set a follow-up review date for each case